I am a PhD student at Princeton University, advised by Karthik Narasimhan.

In my work, I study methods, evaluations, abstractions, and risks of language agents.

In my life, I read, travel, rap, and play basketball.

To anyone: give me feedback about anything!

To students: email me to schedule a chat about anything! I commit 30min/week and priortize underrepresented groups.

Recent News

• Sep 2023: Excited to release 🐨CoLLA, a systematic framework for language agents! Summary here.
• Aug 2023: I gave a two-hour talk covering ReAct, Reflexion, ToT, WebShop, InterCode, Collie. Slides here.
• Jul 2023: I enjoyed teaching at Princeton AI4ALL! Coverage here.
• Jul 2023: I wrote a blog post with Karthik about opportunities and risks of language agents! Comment here.
• Apr 2023: I attended LangChain’s Agent webinar. Summary here.

Recent Papers

• Embers of Autoregression: Understanding Large Language Models Through the Problem They are Trained to Solve
  R. Thomas McCoy, Shunyu Yao, Dan Friedman, Matthew Hardy, Thomas L. Griffiths
  paper | repo | tweet

• Cognitive Architectures for Language Agents
  Shunyu Yao*, Theodore Sumers*, Karthik Narasimhan, Thomas L. Griffiths
  paper | repo | tweet

• Referral Augmentation for Zero-Shot Information Retrieval
  Michael Tang, Shunyu Yao, John Yang, Karthik Narasimhan
  paper | repo | tweet

• Collie: Systematic Construction of Constrained Text Generation Tasks
  Shunyu Yao*, Howard Chen*, Austin Wang*, Runzhe Yang*, Karthik Narasimhan
  paper | repo | tweet | project

• InterCode: Standardizing and Benchmarking Interactive Coding with Execution Feedback
  John Yang, Akshara Prabhakar, Karthik Narasimhan, Shunyu Yao
  NeurIPS 2023 Datasets and Benchmarks Track
  paper | repo | tweet | project

• Reflexion: Language Agents with Verbal Reinforcement Learning
  Noah Shinn, Federico Cassano, Beck Labash, Ashwin Gopinath, Karthik Narasimhan, Shunyu Yao
  NeurIPS 2023
  paper | repo | tweet

• Tree of Thoughts: Deliberate Problem Solving with Large Language Models
  Shunyu Yao, Dian Yu, Jeffrey Zhao, Izhak Shafran, Thomas L. Griffiths, Yuan Cao, Karthik Narasimhan
  NeurIPS 2023 (Oral)
  paper | repo | tweet

• ReAct: Synergizing Reasoning and Acting in Language Models
  Shunyu Yao, Jeffrey Zhao, Dian Yu, Nan Du, Izhak Shafran, Karthik Narasimhan, Yuan Cao
  ICLR 2023 (Oral)
  paper | repo | tweet | project | Google AI blogpost

• WebShop: Towards Scalable Real-World Web Interaction with Grounded Language Agents
  Shunyu Yao*, Howard Chen*, John Yang, Karthik Narasimhan
  NeurIPS 2022
  paper | repo | tweet | project | demo | Quanta Magazine

Recent Talks

• Cognitive Architectures for Language Agents
  • MiraclePlus, Remote, Sep 2023
  • Cognitive Science Colloquium, Princeton, Sep 2023
• From Language Models to Language Agents
  • Tsinghua University, Remote, Aug 2023
• Tree of Thoughts: System 2 Reasoning, Tree Search, and Language Agents
  • Google DeepMind, Remote, Jul 2023
  • Allen Institute for AI, Seattle, Aug 2023
• Towards Autonomous Language Agents
  • DeepMind, London, Apr 2023
  • EPFL NLP Group, Lausanne, Mar 2023
  • Berekley NLP Group, Berekley, Mar 2023
  • OpenAI, San Francisco, Mar 2023
  • Adept AI, San Francisco, Mar 2023
• Re-thinking Reinforcement Learning in the Era of Large Language Models
  • Jane Street, New York, Apr 2023
• ReAct: Synergizing Reasoning and Acting in Language Models
  • ICLR, Kigali, May 2023
  • EMNLP EvoNLP & NILLI Workshops, Abu Dhabi, Dec 2022
  • NeurIPS FMDM Workshop, New Orleans, Nov 2022
• Towards Human-like Language Models
  • Cognitive Science Colloquium, Princeton, Nov 2022
• WebShop: Towards Scalable Real-World Web Interaction with Grounded Language Agents
  • University of Washington, Seattle, Jul 2022

Recent Readings

• Set Theory (John Burgess)
• The Computer and the Brain (John von Neumann)
• Automata Studies (Editted by C.E. Shannon and J. McCarthy)
• Team of Rivals (Doris Goodwin)
• The Linguistics Wars (Randy Harris)
• What Babies Know (Elizabeth Spelke)

(last updated: Sep 2023)